leavelet/cutlass
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

v4.4.1 update (#3079)

Browse Source
This commit is contained in:
Junkai-Wu
2026-02-28 02:59:21 +08:00
committed by GitHub
parent c651d660d2
commit 3bb6e28d3c
13 changed files with 92 additions and 23 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
include/cutlass/gemm/collective/sm90_mma_array_tma_gmma_ss_warpspecialized_fp8_blockwise_scaling.hpp
View File

@@ -212,7 +212,8 @@ struct CollectiveMma<
static_assert(cute::is_same_v<ElementAccumulator, ElementBlockScale>,
"ElementAccumulator and ElementBlockScale should be same datatype");
using NumSplitsM = cute::C<get<0>(TileShape_{}) / 128>;
// For TileShapeM < 128, NumSplitsM should be 1
using NumSplitsM = cute::conditional_t<get<0>(TileShape_{}) < _128{}, _1, cute::C<get<0>(TileShape_{}) / 128>>;
static_assert(NumSplitsM{} == 1 || NumSplitsM{} == 2);
struct SharedStorage {

2
include/cutlass/version.h
View File

@@ -36,7 +36,7 @@
#define CUTLASS_MAJOR 4
#define CUTLASS_MINOR 4
#define CUTLASS_PATCH 0
#define CUTLASS_PATCH 1
#ifdef CUTLASS_VERSIONS_GENERATED
#include "cutlass/version_extended.h"